Privacy without Noisy Gradients: Slicing Mechanism for Generative Model Training

Open in new window